Innovative Floor Truss Top Chord for Achieving Long Spans

Steel and concrete composite systems are generally used as major structural components in multi-storey buildings. Composite construction in buildings is more popular with steel decking since it serves as a working platform to support the construction loads and also as permanent formwork for concrete. To achieve large column free spans (in the range of 8m-12m), as often demanded for multistorey office buildings, "steel and concrete composite floor trusses" may form economical solutions since they provide the facility to accommodate various service ducts within the structural zone. The concept of introducing a concrete filled steel tube (CFST), instead of the conventional open flanged steel section, as the top chord of these floor trusses has been discussed. However, the viability of this new concept should be ensured by experimental evidence on the longitudinal shear transfer capacity at the composite stage. This paper discusses the experimental results of a series of push-off tests conducted on CFST embedded composite slab panels. The effect of providing different concrete top cover and effect of different concrete strength have been investigated, and the results are compared with existing practice related to headed shear studs.


Background
Composite construction using steel and concrete has been used since the early 1920s. It gained widespread use in bridges in the 1950s and in building in the 1960s in the world. The reason for that is structural designers have been aware of the advantages of composite construction, such as saving of steel, reduction of overall structural depth, and the increase in floor stiffness and load capacity.
Composite construction in buildings has become more popular with the profiled steel sheeting since it serves as a working platform to support the construction loads and permanent formwork for the concrete. This eliminates the need for traditional, temporary forms and falsework. Also the sheets are suitably shaped to ensure proper bond with the concrete, the sheeting can provide all or part of the main tension reinforcement in the slab.
Steel and concrete composite systems are generally used as major structural components in multi-storey buildings. Therefore structural arrangement of "floors" is particularly important. Several different configurations of composite floor systems are in use worldwide for long spans, namely, composite stub-girders, slim floor systems, composite trusses, composite beams with web openings in the steel beam, etc.
A "composite truss" is a steel truss, the top .chord of which is designed to act compositely with a concrete slab right above it. To achieve large column free span (in the range of 8m 1 2m), as often demanded for multy-storey office buildings, "steel concrete composite trusses" may form an economical solution since they provide the facility to accommodate various service ducts within the structural zone (i.e. these could be passed through the openings in the truss) which would otherwise have to be placed underneath it.
In this type of construction, the bare steel truss is generally expected to withstand the construction stage loads until the composite action develops in the top chord when the concrete is hardened. Consequently the size of the steel top chord member is governed by the construction stage loading (non-composite action). This means the composite truss contains The property of a shear connector of most relevance to design is the relationship between the shear force transmitted and, the corresponding slip at the interface. This loadslip curve should ideally be found from tests on full-scale composite beams, but in practice a simpler "push-off specimen" is used (see Figure  1). Failure mode of shear connectors in composite slabs could also be found from the push-off tests.
To design composite trusses with CFSTs in top chords, further experimental evidence is required on the shear carrying capacity at the composite stage. Two configurations using 114mm diameter pipes, as the truss top chord were proposed for this study. a) In configuration 1, the top chord was an embedded concrete filled steel tube (CFST) in composite slab and in composite slab it acts as a continuous circular shear connector.
The most common way used to evaluate the shear connector strength and the behavior is the push-off test. Push-off tests have been used as early as the 1930s and were used to predict the strength of studs in solid slabs as described above.
The longitudinal shear force in composite beams is transferred across the steel flange/ concrete interface at a discrete number of points, by the dowel action of the individual shear connectors.
If the concrete slab fails to resist the longitudinal shear stresses produced by connectors, longitudinal cracking along the line of the beam may occur. This leads to a loss of interaction between the steel beam and the concrete compression flange, and a drastic reduction in the moment capacity of the composite section [3].  [2]. more than adequate amount of structural steel fixed to its top chord for the serviceability and ultimate design states. An economical design may hence be achieved by introducing lternative means for the truss top chord, which is to reduce structural steel. Instead of the onventional open flanged steel section, as the top chord of these floor trusses, one such lternative is to use a concrete filled steel tube as described in this study. ' The use of hollow steel tubes filled with concrete has become wide spread in the past few decades. This is mainly due to their properties uch as high strength, high ductility, and large nergy absorption capacity. In this type of omposite trusses the uncertainty is with the hear connection than its compressive strength apacity as a top chord of the truss. The viability f this concept could be ensured by xperimental evidence on the longitudinal shear arrying capacity at the composite stage.
In conventional composite beams, shear carrying ·• pacity is gained by mechanical connectors (to r sist specially longitudinal shear), the most pular form being welded headed studs. The hear studs are welded to the flange of the steel am, generally through a composite steel deck. A composite slab is cast on top of the deck with th stud, functioning to tie the slab and beam t gether as unit. A composite beam has greater trength and stiffness than if the beam and slab were behaving independently. In concrete filled l el tube (CFST) embedded composite trusses which is suited in this study, the shear carrying , pacity is gained by surface area of concretet el contact. · b) In configuration 2, headed shear studs were used as shear connector and top chord of composite slab was a CFST but it was not embedded to slab.
The proposal was for testing of specimens of the two configurations by varying the clear-cover thickness in the concrete and the concrete grade since effect of those were not well known, The project objectives are as described below.
• To experimentally verify feasibility of the configuration suitable For top chord members of composite trusses • To determine the effect of concrete top cover on shear transfer capacity of new deck slab configurations.
• To determine the effect of compressive strength of concrete, on shear transfer capacity of new deck slab configurations.
• To identify the pattern of shear failure planes for each of new deck slab configuratio.,s.

Literature Review
This review will begin with the strength prediction equations developed for welded headed studs in solid slabs and will include more recent research on welded headed studs in slabs with formed metal deck.
Ollgaard et al [4] performed 48 solid slab pushoff tests. Variables considered were concrete compressive strength, concrete splitting tensile strength, modulus of elasticity of concrete, density of concrete, stud diameter and type of aggregate. The stud tensile strength, slab reinforcement, and geometry were constant for all tests. Stud diameters tested were 5/8 i.1. 1 and 3 / 4 in. Two types of normal weight concrete and three types of lightweight concrete were tested. The failure modes observed were stud shearing, concrete failure, and a combination of the two.
For the concrete failure mode, the lightweight concrete had more and larger cracks in the slabs than did the normal weight concrete. When one pair of connectors was in each slab, all failed by shearing off the studs. The lightweight concrete tended to crush in front of the studs, causing the stud to remain straight when it deformed. When lightweight concrete was used, stud strengths decreased 15 % to 25 % .
The normal weight concrete provided greater restraint of the stud, so more curvature of the stud occurred. Studs in both types of concrete rotated a large amount at the weld. The tests showed that studs in both types of concrete exhibited considerable inelastic deformation before failure as the specimens did not fail suddenly at ultimate load.
The tests showed that the stud strength decreases when the concrete strength decreases considerably. The data indicated that the stud strength is more influenced by the concrete compressive strength and modulus of elasticity than by the concrete splitting tensile strength and density. They also concluded that the concrete properties control at ultimate load, so shear connector tensile strength is not as critical; however, smaller diameter connectors would be more dependent on stud tensile strength, because the concrete forces would not be as great.
Elkelish and Robinson [5] studied six parameters that affect longitudinal cracking of composite beams with metal deck using experimental specimens as well as a finite element analysis. The parameters investigated were type of loading, concrete compressive strength, beam span-to-slab width ratio, thickness of the solid part of the slab, percentage of transverse reinforcement, and the existence of the metal deck. Three loading conditions were used: uniformly distributed load, single point load at mid-span, and two point loads at the third points.
The results showed that a uniformly distributed load causes the longitudinal crack to start at the top of the slab; single point load and two point loads cause the crack to start at the bottom. The initial longitudinal crack is delayed with an increase of the span-to-width ratio, the steel beam yield strength-to-concrete strength ratio, the thickness of the solid part of th slab, and the transverse reinforcement ratio. Welded wire mesh did not increase th r i t nc to initial longitudinal cracking. Th m l I d k helps to The connection resistance (QK) can then be found as r ist cracking when point loads are applied, but ll does not help for a slab that is uniformly I aded.

Methodology
an elastic-plastic type of curve with a yield plateau corresponding to the connector characteristic resistance P,u and to a high ultimate slip capacity su. Eurocode 4 considers that connectors having a characteristic slip capacity higher or equal to 6 mm can be assumed to be ductile, provided that the degree of shear connection is sufficient for the spans of the beam being considered.
In brief, Conventional headed shear connectors' shear carrying capacity increase with concrete grade and concrete top cover. Therefore, in this study different concrete grade and concrete top cover were used to find out shear carrying capacity with two new configurations.
In Sri Lankan construction industry practical concrete grades with slabs are grade 20 and 30. The economical concrete top covers are 20mm and 25mm. Therefore, for each configuration C20, C30, and C45 concrete were used with concrete top cover 25mm to check the effect of concrete grade on shear transfer capacity. Similarly concrete top cover 20mm, 25mm, and 30mm were used with C30 concrete to check effect of concrete top cover on shear transfer capacity. In each case used three replicates were tested to verify the results. All together forty five samples were tested for all three configurations and effect of configuration on shear transfer capacity also checked.
The proposed "push-off test (push test)" has a single span arrangement with only one deck slab specimen, as opposed to "standard push-off tests" described in Standard Codes of Practice [9] [10] for conventional composite arrangements where two identical deck slab specimens to be fixed on either side of the main steel beam. The major reason for that is the proposed configurations are not containing open flange steel beam sections. A considerable amount of material saving is also expected by the proposed arrangement. A similar arrangement has been successfully used for push off tests in the past [3]. The testing arrangement is illustrated in Figure 3 In almost all of the tests, surface cracks appeared along with separation of the concrete from the d ck just before ultimate load was reached. After ultimate load, the slabs were seen to ride ov r the sheeting and cause extensive profile distortion. Wedge-shaped failure cones, not pyramidal-shaped cones as suggested by I law kins and Mitchell [7], occurred around the luds in all of the tests. This mechanism has b en found to occur in a composite beam test.  The load was first applied in increments up to 40% of the expected failure load and then cycled 25 times between 5% and 40% of expected failure load. Subsequent load increments were then imposed such that failure did not occur in less than 15 minutes. The longitudinal slip between each concrete slab and concrete filled steel tube was measured continuously during loading or at each load increment. As close as possible to each group of connectors, the transverse separation between the steel section and each slab was measured.

Experimental results
During the analysis, mean value of each variable was considered and outliers were neglected.
All push-off test results in each configuration are presented in Table 4.1, and 4.2 with average concrete cube strength of four samples. In each configuration concrete cube strength, and concrete top cover were used as variables.
Two layers of steel reinforcement (R6@200c/ c) were placed top of the steel tube. Mortar cover blo,cks were used to support the reinforcement. Silicon was applied to fill the openings. All slabs were cast horizontally. After the concrete was placed in the forms, it was vibrated with a mechanical vibrator. verticality of the deck slab specimens during loading, as shown in Fig. 3.l(b).
The concrete filled steel tube was welded to steel plate prior to concrete filling. Then it was filled by concrete and concrete test cubes were Last at the same time. Concrete test cubes were placed without curing (to check behaviour of compressive strength of concrete, in sidz the steel tube), but with proper covering.
In this configuration shear transfer capacity of composite slab relies on contact between steel tube and concrete. In other words friction force between steel and concrete enhances the shear transfer capacity of composite slab in configuration 1 which is different from conventional headed shear studs in configuration 2.

Test Specimens
All push off slab specimens were constructed using wooden forms. Each specimen consisted of concrete filled steel tubes, each of which was welded to a 200mm wide x 950mm long x 9mm thick steel plate. Two "S" shape flashings (25mm x 50mm x 20mm) riveted to two steel profile sheets, were riveted to steel plate as shown in Figure 3.2(a).

TOP VIEW
The specimens were covered and moist-Lured for seven days, at which time the forms were removed. Concrete test cubes were . cast along with the specimens and cured similarly. The push-off specimens were tested 28 days after being cast. At that time both concrete test cube samples were tested.

Configuration 2
In this configuration conventional headed shear studs (19mm dia., 80mm height) were used to improve the shear transfer capacity of composite slab (see        With configuration 1 and 2 the failure load increased 16% to 45% and 4% to 33% with increase of concrete strength respectively. Configuration 1 shear carrying capacity was nearly four times the configuration 2 capacity.      For comparison average test results were used.

Effect of Concrete Strength on Shear Connectors
The effect of concrete strength on the strength of shear connectors was checked in this study and it was checked for configuration 1, and 2 as shown in Table 4.3, 4.4 and Figure 4.1 .

58
The effect of concrete top cover on the strength of shear connectors was checked in this study and it was checked for configuration 1, and 2 as shown in Table 4.5, 4.6 and Figure 4.2.
With configuration 1 and 2 the failure load increased 16% to 21 % and 16% to 26% with increase of concrete top cover respectively. Configuration 1 shear carrying capacity was nearly four times the configuration 2 capacity.
.  [11,12]. With configuration 2, 4%-33% of increase in shear carrying capacity was achieved by increasing concrete cube strength from grade 20 to grade 45. Also 16%-26% of increase in shear carrying capacity was achieved by increasing concrete top cover from 20mm to 30mm.
With configuration 1, 16%-45% of increase in shear carrying capacity was achieved by increasing concrete cube strength from grade 20 to grade 45. Also16%-21 % of increase in shear carrying capacity was achieved by increasing concrete top cover from 20mm to 30mm.
The connector type for configuration 1 was non ductile and it was ductile with configuration 2.
The shear failure pattern propagated at minimum top cover at configuration 1. One longitudinal crack was seen at top of the steel tube in configuration 1, and Shear cone failure was seen with configuration 2.
Therefore Configuration 1 is more suitable for composite floor truss top chord compared with configura tion2. Where, n = 0.44 for Configuration 1 n = 0.36 for Configuration 2 I• r configuration 2 that is shear connection with h iaded shear studs the plots against slip, and Iailure load were showed Figure 2.l(a) type I, haviour. But configuration 1 was showed l·igure 2.1 (b) type behaviour. As a result of th se, it can state that configuration 2 shear , onnection is ductile and a configuration 1 has n n-ductile shear connection (see Figure 4.3).
. 4 Shear Failure Pattern with each Configuration ' l'he shear failure pattern of configuration 1 is hown in Figure 4.4; the longitudinal crack was I ropagated on top of the embedded steel tube. The failure of the composite slab is mainly due t loss of friction force at steel concrete interface.
The shear failure pattern of configuration 2 is , hown in Figure 4.5, longitudinal crack, and transverse cracks were propagated on top of the h aded shear studs. The failure of the composite I b is mainly due to pulling out of stud, and it I known as shear cone failure (Wedge cone ( ilures were seen).
Prediction failure load was compared with xperimental failure load as shown in Figure 5